mIoU: 65.86%mAP: 86.56%
Papers with Code · CC BY-SA 4.0
Scene Parsing
Scene parsing is to segment and parse an image into different image regions associated with semantic categories, such as sky, road, person, and bed. MIT Description
논문
mIoU: 68.26%mAP: 79.65%
3DIoU: 79.36
3DIoU: 76.33Corner Error: 1.04Pixel Error: 2.7
3DIoU: 84.93Corner Error: 0.67Pixel Error: 1.93
3DIoU: 79.79Corner Error: 0.71Pixel Error: 2.39
3DIoU: 81.2Corner Error: 0.78Pixel Error: 2.37
Recall@50: 13.73Recall@100: 16.88mAP: 20.08
3DIoU: 84.66Corner Error: 0.60Pixel Error: 1.97
Recall@50: 11.56Recall@100: 14.13mAP: 18.9
3DIoU: 82.43Corner Error: 0.70Pixel Error: 2.25
3DIoU: 78.79%
3DIoU: 79.88
3DIoU: 83.77